IN-91-09-4 -- Passive Solar Design for Windows Using a Neural Network PDF

In passive solar systems, the large south-facing glazed openings designed for winter heating are often troublesome in summer cooling. One of the effective means of reducing summer solar energy gain through the passive openings is the installation of overhangs. This paper uses a neural network approach to predict how a fixed overhang over a south-facing window can be designed so that it shades the window completely in midsummer and leaves it totally unshaded in midwinter. The predictions are made for 28 N latitude to 44 N latitude, the spread for the lower 48 states. Window heights of 3 to 5 feet, overhang sizes of 1 to 3 feet, and a height at which the overhang has to be installed from the top of the window of 1 to 2 feet. All predictions are based on an average of five hours daily, i.e., starting at 10.00 a.m. and ending at 3.00 p.m.

KEYWORDS: passive, solar heating, windows, solar radiation, shades, summer, winter, calculating, designing, solar heat gain, expert systems